regional production network, asian trade …1 regional production network, asian trade integration,...
TRANSCRIPT
1
Regional production network, Asian trade integration, and optimal
monetary policy coordination
Chin-Yoong Wong
Department of Economics, Faculty of Business and Finance, Universiti Tunku Abdul Rahman. Correspondence author. Email address: [email protected] Yoke-Kee Eng Department of Economics, Faculty of Business and Finance, Universiti Tunku Abdul Rahman. Muzafar Shah Habibullah Department of Economics, Faculty of Economics and Management, Universiti Putra Malaysia
Abstract Through the lens of Bayesian-estimated two-region New Keynesian model with vertical and processing trade, this paper provides an unified framework to shed lights on the magnitude of production sharing in East and Southeast Asia, the measure of vertical specialization between China, East, and Southeast Asia, a propagation mechanism demonstrating how these economies are interdependent, and the welfare-maximizing monetary policy rule. JEL classification: C11, E52, E61, F15, F41, F42
[VERY PRELIMINARY DRAFT]
[FIRST VERSION: APR 6, 2012]
2
1. Introduction
One of the defining characteristics of today’s international trade is the heavy
flow of parts and components across countries. Underlying the fast-growing world
trade since trade barrier reduction in 1980s, which brings about the integration of
majority developing countries into world economy, is the reorganization of production
structure. Productions are vertically sliced and fragmented across countries with
different factor endowments, and as a result, multiple back-and-forth trades in
intermediate goods are generated before a final product is assembled (Feenstra
1998). And this vertical intra-industry trade under the umbrella of international
production network has accounted for an increasing share of total trade (Hummels et
al. 2001; Yi 2003).
While the formation of production network is equally observed in the rest of the
world (see, for instance, Koopsman et al. 2010), the depth and complexity of vertical
production network and trade in East Asia is unrivalled. In a study of 79 countries,
over 121 categories of goods within the period of 1967-2005, Amador and Cabral
(2009) show that out of top ten vertically most specialized countries, eight are
located in East Asia. Of 22 countries in East, Southeast, South, and Central Asia in
2003, Sawyer et al. (2010) document that Southeast Asian countries and the high-
income economies in East Asia exhibit the highest level of intra-product trade,
followed closely by China (see, also, Athukorala and Yamashita, 2009).
In view of the uniqueness and importance of production sharing in this region,
this paper makes an effort to provide a perspective on the magnitude of production
fragmentation in East and Southeast Asia, a measure of vertical specialization
between China, East, and Southeast Asia, a propagation mechanism demonstrating
how these economies are interdependent, and a framework for welfare-maximizing
3
monetary policy rules through the lens of a Bayesian estimated two-region New
Keynesian model with vertical and processing trade.
Overall, East and Southeast Asia are tied closely through vertical trade in
intermediate inputs. The integration of China into world trading system, however, has
restructured the regional production nexus in the sense that East Asia has nearly
completely specialized in midstream production while China in downstream
production. The corollary is strong complementarity between East Asia and China
embodied in the sequential vertical and processing trade. But such complementary
relationship is weak between China and Southeast Asia. Despite the fact that
Southeast Asia still engages in midstream production and thus vertical trade with
China, its importance in total trade has been declining while seeing rising share of
processing trade in total exports to China. This simply implies that competitive
relationship between China and Southeast Asia has been intensifying since China’s
WTO accession. Following the tradition of New Keynesian literature as in Woodford
(2003), we use a quadratic approximation of the utility-based welfare criterion around
the non-distorted global maximum utility to characterize the cooperative allocation for
open economies engaging heavily in vertical and processing trade.
The following discussion is organized into four sections. Section 2 lays out a
two-region New Keynesian model with three processing stages to explicitly
incorporate vertical and processing trade. Section 3 details the Bayesian method and
data used in evaluating the trade integration between East Asian economies.
Throughout, we shed lights on the measurement of vertical linkage between China,
East, and Southeast Asia, the source of fluctuations, and the influence of China on
East and Southeast Asia particular in the aftermath of WTO accession. In Section 4,
we probe into the welfare-maximizing operational monetary policy rule based on
estimated models. Of particular interest is whether optimized policy necessitates
4
response toward foreign disturbances in home policy feedback rule. As a further,
whether welfare can be enhanced if there is policy coordination? Section 5
concludes.
2. A macroeconomic model of vertical and processing trade
In this section we lay out a two-region New Keynesian model of vertical and
processing trade (NKVPT) according to Wong and Eng (2011). There are three
processing stages in the NKVPT model. The upstream firms combine labor services
and capital in Cobb-Douglas production technology to produce materials that will be
partially exported abroad for subsequent processing in midstream production.
Speaking differently, home midstream firms would fabricate the imported
intermediate goods in conjunction with local upstream processed materials. A
fraction of remanufactured intermediate goods would then be re-exported for final
assembly in downstream production. The assembled final goods are to be consumed
and invested locally as well as exported to trading partners.
By considering three processing stages, the NKVPT model embraces
simultaneously vertical trade and processing trade. While the former is back-and-
forth trade in intermediate inputs, the latter involves imports of intermediate inputs for
remanufacturing and re-exporting as consumer goods. We believe that such model
design is important in order to more satisfactorily gauge the emergence of China as
“world factory” that final assembles the intermediate inputs absorbed from intra-East
and Southeast Asia for consumer markets in the U.S. and Euro Area on the one
hand, and the role of other Asian countries in importing, manufacturing, and
exporting intermediate goods on the other hand.
5
What follows is the brief technical description of the NKVPT model.
2.1 Upstream firms
A unit mass of competitive firms at upstream production has access to Cobb-
Douglas production technology of Eq. (1) that uses plant-specific capital �������,
� ∈ , and differentiated labor ���� of a variety � ∈ � to produce plant-specific
materials ����� at date t.
����� = ���������������������� (1)
where �� is an first-order autoregressive Hicks-neutral total factor productivity (TFP)
shock. The upstream firms can only alter its capital over time by varying the rate of
investment ����� that comes with a cost S������ �������⁄ �. Capital stock accumulation
evolves according to the form in Mandelman et al. (2011).
����� = �1 − ��������� + u������� �1 − �� �u��� �����!�
u� ���!� " � #� ���!�#��� �����!� − $"�% (2)
where u�� is investment-specific technology (IST) shock, and follows first-order
autoregressive process. The parameter Ψ denotes investment adjustment cost, and
Λ determines how forward-looking the investment decision is. The upstream firm
thus optimally chooses the path of �� and � '= () ����� *�+⁄ ,�-∈� .*�+/ to minimize the
cost of production
01,�������� + 3�� (3)
subject to production function net of investment adjustment cost
Φ�� 5���������������������� − Ψ2 u���� ������� ' 7�������7���� ������� − Λ/� = 09
where �� is Lagrangian multiplier which we define as the real marginal cost for
upstream firm. 3� = () 3�����:+,���� :+,�⁄ ,�-∈� .:+,� �:+,����⁄ refers to the real wage, 01,� is
the real return on capital, and ;<,� denotes the wage elasticity of the demand for
6
labor � . The optimal demand for labor of varieties � , total capital and labor, and
investment decision are given by
���� = �=��-�=� "�:+,� � (4)
������� = > �?@,�A B� ����� (5)
� = � �=�" �1 − B�Φ� ����� (6)
Φ1C � 7C��C���7C−1� �C−1��� − Λ" = Φ1C+1 D�7C+1� �C+1���7C��C��� − Λ" �7C+1� �C+1���7C��C��� " − �� �7C+1� �C+1���7C��C��� − Λ"�E (7)
For the sake of simplicity, we assume that the market for upstream goods is tightly
competitive. The elasticity of substitution between varieties thus is close to infinity,
and as a consequence, price approximates real marginal cost. The pricing decision
is further assumed to be symmetry across manufacturing plants.
2.2 Midstream firms
A mass continuum of midstream monopolistically competitive firm j, � ∈ ,
imports upstream goods F�G,�! of plant j for remanufacture. In combination with local
inputs �I,�!, the midstream firm j uses CES production technology as in Eq. (8) to
produce intermediate goods for subsequent processing.
��! = J�1 − K��� L⁄ M �I,�! N�L��� L⁄ + K�� L⁄ MF�G,�! N�L��� L⁄ OL �L���⁄ (8)
where
F�G,�! = 'P F�G,�! ����:Q���� :Q�⁄ ,�!∈R /:Q�:Q���
and
�I,�! = 'P �I,�! ����:Q���� :Q�⁄ ,�!∈R /:Q�:Q���
7
The demand function for the j varieties of F�G,�! ���is MS�G,�! ��� S�G,�!T N�:Q�F�G,�!, and of
�I,�! ��� is MS�I,�! ��� S�I,�!T N�:Q� �I,�!. S�I,�!
and S�G,�!, respectively, is the home price of
local and imported materials. ;�� > 1 is the time-varying demand elasticity. The
parameter K� indicates the share of imported parts and components in midstream
production, and the parameter V > 0 denotes the elasticity of substitution between
home and imported intermediate inputs. The optimal demand function for home and
imported materials can be derived in the following form
�I,�! = �1 − K�� 'W�X,�YW��Y /�L ��! (9)
F�G,�! = K� 'W�Z,�YW��Y /�L ��! (10)
where
S��! = J�1 − K��MS�I,�! N��L + K�MS�G,�! N��LO� ���L�⁄ (11)
S��! is the flexible producer price for midstream production.
2.3 Downstream firms
Lastly at downstream production, a continuum of monopolistically competitive
final-good producers j of measure J combines a variety of home �I,�! and imported
intermediate goods F�G,�! using the following CES technology to produce consumer
goods.
[�! = J�1 − K[�� L⁄ M �I,�! N�L��� L⁄ + K[� L⁄ MF�G,�! N�L��� L⁄ OL �L���⁄ (12)
where
8
�I,�! = 'P �I,�! ����:\���� :\�⁄ ,�!∈R /:\�:\���
F�G,�! = 'P F�G,�! ����:\���� :\�⁄ ,�!∈R /:\�:\���
The parameter K[ denotes the share of imported intermediate inputs in final-good
production. Solving for downstream firms’ cost minimization problem, as in the case
of midstream firms, gives us the demand schedules for the varieties and for home
and imported intermediate goods shown in Eqs. (13) – (16), respectively.
�I,�! ��� = 'WQX,�Y �!�WQX,�Y /�:\� �I,�!
(13)
F�G,�! ��� = 'WQZ,�Y �!�WQZ,�Y /�:\� F�G,�!
(14)
�I,�! = �1 − K[� 'WQX,�YWQ�Y /�L [�! (15)
F�G,�! = K[ 'WQZ,�YWQ�Y /�L [�! (16)
S��! is the producer price index for final-good producers:
S��! = (�1 − K[�MS�I,�! N��L + K[MS�G,�! N��L.� ���L�⁄ (17)
2.4 Optimal pricing decision with U.S dollar pricing in export
Pricing decision is assumed to be time dependent. The ability of domestic firms
at midstream and downstream production to re-optimize the price is subject to the
signal received at probability 1 − ]W^ , for _ = 2,3 . Firm � that receives the signal
chooses ℙ^I,� to maximize the expected discounted profits b�Π� for sales in home
9
market and ℙ^I,�d for export market. For home market, the pricing decision is
formulated as
b�Π�efgh = b� ∑ �]W^j�-Λ�l- JℙmX,�no�!��pqm,�noWm,�no O JℙmX,�no�!�WmX,�no O�:m,� I,�l-���r-st (18)
Contrary to producer-currency pricing decision in the typical New Keynesian
model, or local-currency pricing in the New Open-Economy model, we consider U.S.
dollar pricing (DP) strategy in exports. This assumption is apparently coherent with
the fact that international trade is largely denominated in the U.S dollar (Goldberg
and Tille, 2008). The variability of exchange rates between local currency and the
U.S. dollar uId,� will not pass through into foreign price of home export, but rather, it
will pass through into local-currency denominated export earnings. Firm’s expected
export profit in home currency under DP strategy is thus given by
b�Π�dW = b� ∑ �]W^∗ j�-Λ�l- JwXx,�ℙmX,�nox �!��pqm,�noWm,�no O ywZx,�ℙmX,�nox �!�WmX,�no∗ z�:m,� I,�l-∗ ���r-st (19)
In what follows we assume that all firms are symmetric in price setting.
Firms allowed for price re-optimization will reset their log-linearized price ℙ{^I,�^h|
to approximate the optimal reset price derived from Eqs. (12) and (13), respectively,
for home and export market. The remaining firms that do not receive signal for re-
optimization will stick to last-period price, out of which a fraction of them }W^ will
index to last-period inflation. The corresponding inflation dynamics of PPI (~�I,��, GDP deflator (~[I,��, intermediate export price �~�I,�∗ � and final export prices �~[I,�∗ � can be derived as
~^I,� = � ��m�l��m���m" ~^I,��� + � ��l��m���m" b�~^I,�l� + �M0��� ^,� + �^,�N (20)
~^I,�∗ = > ��m∗�l��m∗ ���m∗ A~^I,���∗ + > ��l��m∗ ���m∗ Ab�~^I,�l�∗ + �∗M0��� ^,� − �Id,� + �^,�∗ N (21)
10
where
� = �1 − ]W^��1 − ]W^j� M]W^�1 + ]W^j}W^�N⁄ ,
�∗ = M1 − ]�^∗ NM1 − ]�^∗ jN �]�^∗ M1 + ]�^∗ j}�∗ N"T , and �^,� is an i.i.d price markup shock
for _ = 2,3. 0��� ^,� is the log-deviation of real marginal cost, in which 0��� �,� = ��,� and 0��� [,� = ��,�. 2.5 Household
Consider a continuum of infinitely-lived households, represented and indexed
by i ∈ �0,1�,who possess the utility function of
� = b� D∑ j�u�q J�q�o�I�������� − u�< �<�o��n�
�l� Or�st E (22)
where
��- = (�}�� �⁄ M�I,�- N����� �⁄ + �1 − }�� �⁄ M�G,�- N����� �⁄ .� �����⁄
(23)
u�q and u�< , respectively, is i.i.d preference and labor supply shock. �I,�- �= �) �I,�- ����:���� :�⁄ ,�-∈� " ������% and �G,�- �= �) �G,�- ����:���� :�⁄ ,�-∈� " ������% are the
composite varieties of home and imported consumer goods. ���= �����- � indicates
external habit formation in which b is the parameter that governs the extent of habit
persistence. 0 < j < 1 refers to subjective discount factor, � measures the
coefficient of relative risk aversion, and the reciprocal of � indicates the wage
elasticity of labor supply. The parameter � > 1 denotes the elasticity of substitution
between home and imported consumer goods. The parameter }measures home
bias. Household i’s constrained optimization problem can be illustrated as utility
maximization of Eq. (22) subject to Eq. (23) and the following flow budget constraint
�� + �wXx,�W���" ���∗��x" + ��W��� + �� = 3� � + Π� + 01,����� + �wXx,�����∗ l����W� " (24)
11
where SI,� and SG,� , respectively, denotes domestic price of home and imported
consumer goods. Household facing exchange-rate risk � in foreign asset market
has access only to imperfect international asset market. Note that the foreign bond
¡�∗ is denominated in U.S. dollar. Thus, the nominal exchange rate between home
currency and the U.S. dollar is considered. Solving for the utility maximization
problem gives us the optimal demand schedules for varieties and composite
varieties as in Eqs. (25) – (28), marginal rate of substitution between works and
consumption in Eq. (29), intertemporal substitution of consumption in Eq. (30), and
uncovered interest rate parity in Eq. (31).
�I,�- ��� = >WX,��-�WX,� A�:� �I,�- (25)
�G,�- ��� = >WZ,��-�WZ,� A�:� �G,�- (26)
�I,�- = } �WX,�W� "�� ��- (27)
�G,�- = } �WZ,�W� "�� ��- (28)
M�-N�M��- − �����- N�u�< = 3�p�w (29)
Mq�o�¢q���o N��W� u�q = j�1 + 0�� M£�q�n�o �¢q�oN��
£�W�n� b�u�l�q (30)
uId,� = b�uId,�l� ��l?�¤¥�l?� " � (31)
S� is the utility-based consumer price index (CPI).
S� = ¦}SI,���� + �1 − }�SG,����§� �����⁄ (32)
12
Since household is a monopoly supplier of differentiated services, nominal wage is
set in Calvo-style, which results in nominal wage inflation dynamics as what follows:
~�| = ¨ �©�l�©��©ª ~���| + ¨ ��l�©��©ªb�~�l�| + �|�«¬�p�w − «¬� + u�|� (33)
where �| = ����©�����©���©��l�©��©� . ]| denotes wage stickiness, and }| is wage indexation.
u�=is i.i.d wage markup shock.
2.6 Trade balance, aggregate demand, and monetary policy
We define trade balance as the balance between aggregate f.o.b exports and
aggregate c.i.f imports.
� = �I,�∗ + P �I,�∗!∈R ���,� + P [I,�∗
!∈R ���,� − F�G,� − P F�G,����!∈R ,� − P �G,����-∈� ,� (34)
The value added of the economy (GDP) can be defined as
®� = �� + �� + � (35)
The model is closed by considering a general form of monetary policy reaction as
below:
0� = ¯�0��� + �1 − ¯��M0� + °±~qW�,� + °®®{� + °∆w∆�Id,�N + u�� (36)
where 0� �= u�q + ��7�� + ³��" is the natural rate of interest, ¯� measures the interest
rate persistence, °±, °®,and °∆w , respectively, indicates central bank’s
responsiveness toward variability in CPI inflation, aggregate demand variability, and
rate of change in nominal exchange rates between home currency and U.S. dollar.
u�� refers to i.i.d white noise to the conduct of monetary policy.
13
3. Evaluating the trade integration between East and Southeast Asia
3.1 A Bayesian estimation
We take the model to the data using Bayesian method. As the literature on
Bayesian estimation and evaluation has been growing tremendously, its estimation
procedure is briefly sketched here. The procedure is principally built around the
likelihood function of the data derived from the model in conjunction with the prior
belief on the probability distribution of the parameters. Bayesian estimation is thus
about finding a set of parameters that maximizes the posteriors (see, for instance,
Fernandez-Villaverde, 2010; Schorfheide, 2011).
According to Bayes’s theorem, the posterior distribution of model parameters
µ�¶|Υ,ℳ� is formed by combining the likelihood function µ�Υ|¶,ℳ� and prior
density µ�¶,ℳ� in following manner:
µ�¶|Υ,ℳ� = µ�º|¶,ℳ�µ�¶,ℳ�µ�º,ℳ� (37)
where µ�Υ,ℳ� is the marginal density of the data, given a specific model:
µ�Υ,ℳ� = ) µ�Υ|¶,ℳ�µ�¶,ℳ�¶ ,¶ (38)
Suppose that the marginal density of the data is a constant or equals certain
parameter, the posterior kernel can be derived from the numerator of the posterior
density
»�¶|Υ,ℳ� ≡ µ�¶|Υ,ℳ� ∝ µ�Υ|¶,ℳ�µ�¶,ℳ� (39)
where ∝ denotes proportionality. Posterior kernel is simulated to generate draws
using Markov Chain Monte Carlo (MCMC) method. The resultant findings provide the
point estimates, standard deviations and probability density region.
14
3.2 Data and priors
In this paper we are going to study nine East and Southeast Asian economies,
including Japan, the Republic of Korea, Hong Kong, Taiwan, Singapore, Malaysia,
Thailand, Indonesia, and the Philippines, in addition to China. The regional vertical
production and trade link is systematically formed as the consequence of the
massive outflow of vertical foreign direct investment from Japan in the aftermath of
Plaza Accord and subsequently other first-tier Newly Industrializing Economies to
Southeast Asia, and to China. In view of this, we focus on the year 1987 onward and
categorize the nine Asian economies, besides China (CN), into developing
Southeast Asian economies (SEA4), which consists of Indonesia, Malaysia,
Philippines, and Thailand, and advanced East Asian economies (EA5) for the rest.
There are three two-region models to be estimated: SEA4-EA5 model, CN-EA5
model, and CN-SEA4 model. The name that appears first is treated as home region,
while the following as foreign region.
We use 19 macroeconomic observable series in estimating the NKVPT model,
which include real GDP, real consumption, real investment, labor force, nominal
interest rate, nominal exchange rates between home currency and the U.S. dollar,
PPI inflation, GDP deflator inflation, and CPI inflation for SEA4, EA5, and China in
two-region setting, and the U.S. federal funds rate. All the quantity variables are
deflated by respective deflators, and all data, except for the rates of inflation and
interest, are logged and de-trended using Hodrik-Prescott Filter with a smoothing
parameter of � = 1600. We then construct the trade-weighted cyclical observable
series for SEA4 and EA5 using time-varying fraction of national total trade over
aggregate regional trade. We lastly decompose the constructed series into two
subsamples: one from 1987Q1 to 2000Q4 and another from 2001Q1 to 2008Q4.
This is to shed lights on the impact of China’s WTO accession and the subsequent
15
emergence of China as the center of final assembly in the regional production
network on East and Southeast Asian economies.
We assume symmetric priors for estimation. Nonetheless, we allow for different
posteriors for price indexation and stickiness, share of imported materials, home bias
in consumption, monetary policy reaction functions, and standard deviation of
structural shocks. We use Dynare 4.2.5 algorithms for model estimation, and adjust
the number of Markov chains to ensure that estimates for mean and standard
deviation across the Markov chains are satisfactorily consistent.
Because of the lack of previous studies on the Bayesian estimation on the
Asian economies, we are bound to the principle of allowing equal probability for all
potential parameter values within the theoretically coherent range when the true
value is uncertain. As such, priors for the share of imported material at both
midstream and downstream production, price stickiness, and the standard deviation
of shocks are assumed to be in uniform distribution with a range of [0,1]. The priors
for efficient shock persistence are in beta distribution, while the parameters in
monetary policy reaction function, which theoretically must have positive values, are
in gamma distribution with prior means following the standard assumption. The priors
and probability distribution functions are detailed in Table 1.
----- [INSERT TABLE 1 HERE] -----
3.3 Measuring the vertical linkage
Tables 2 to 4 report the estimates of mode, mean and probability interval for
selective interesting parameters in SEA4-EA5 model, CN-EA5 model, and CN-SEA4
model, respectively. All structural shocks and parameters are nicely identified as
evidenced by the proximity between posterior mode and mean which falls within the
16
90% probability interval1. Figure 1 compares the autocorrelations shown in the data
with the one generated from the two-region models, respectively shown in (a)
throughout (c). Evidently, the model is fit in terms of replicating the dynamics shown
in the data.
Due to space constraint, we only pay attention to the share of imported
materials in productions. Table 2 shows that SEA4 and EA5 have been tightly
bonded since 1987 in the sense that the productions in both regions rely heavily on
the intermediate inputs imported from each other. For instance, over the first
subsample periods, the mean share of materials imported from EA5 for midstream
production in SEA4 is 64.9%. The share of materials in EA5 production imported
from SEA4 is even higher. Unsurprisingly, such a strong trade in intermediate inputs
can also be seen in the interaction between CN and EA5, and SEA4.
--- [INSERT TABLES 2 to 4 HERE] ---
To probe into the degree of vertical and sequential trade linkage among these
regions, we construct an index of vertical specialization in the spirit of Hummels et al.
(2001)2. In particular, the degree of vertical specialization of total export for country � over production ℎ is measured by
°u- = ∑ ÀÁ£Â�f?�ÁÁ∑ £Â�f?�ÁÁ (40)
which can be decomposed into vertical trade and processing trade
1 Detailed illustration on the identification of each parameter and shock are available upon request.
2 As defined in Hummels et al. (2001), vertical specialization occurs when (a) a good is produced in two or more
sequential stages; (b) two or more countries provide value-added during the production of the goods; (c) at least
one country must use imported inputs in its stage of the production process, and some of the resulting output
must be exported. Koopman et al. (2010) further classify the value chains into four different categories: (a) final
goods export; (b) intermediate exports that are transformed into final goods and absorbed by the direct importer;
(c) intermediate inputs that are used to produce other intermediates and sent to a third country for further
processing; and (d) intermediate inputs that are transformed into final goods and exported to a third country for
consumption. In this respect, the NKVPT model, which features two-region setting with three sequential
production stages, has an advantage of measuring the vertical specialization embodied in Hummels et al. (2001),
which is conceptually equivalent to the first two categories in Koopman et al. (2010), while capturing the back-
and-forth trade in intermediates as in the category (c) in Koopman et al. (2010). However, the limitation of a
two-region setting is obvious: it cannot measure the vertical and processing trade involving third country.
17
°u- = K� ' bÃ�Ä0C�∑ bÃ�Ä0Ce[� /ÅÆÆÆÆÇÆÆÆÆÈÉh?�-ÊËÌ�?ËÍh
+ K[ ' bÃ�Ä0C[∑ bÃ�Ä0Ce[� /ÅÆÆÆÆÇÆÆÆÆÈ�?fÊhÎÎ-^Ï�?ËÍh
Vertical trade refers to back-and-forth trade in intermediate inputs, that is, importing
intermediate inputs for remanufacturing and reexporting as intermediate input for
subsequent fabrication. Processing trade belongs to the type of trade which imports
intermediate inputs from other countries to provide final goods. Such decomposition
is useful in understanding the nature of trade in East Asian economies given its more
roundabout production network. Koopman et al. (2010) show that for much of East
Asia, including Hong Kong, Korea, Taiwan, Malaysia, and Philippines, use imported
intermediate inputs to provide both intermediate inputs and final goods concurrently
in the world production chain. Meanwhile, Japan lies upstream in the production
chain by providing intermediate inputs, and China lies downstream in the production
chain by using imported intermediate inputs to provide final goods to the world (see,
also, Athukorala and Menon 2010).
Table 5 reports the computation of vertical specialization of total export over
three different estimated models. Based on the estimated share of imported
intermediate inputs in production, and the share of intermediates and final goods in
total exports inferred from Kim et al. (2011)3, Table 5 shows that the share of vertical
specialization in total exports ranges from 0.293 to 0.644 over the period of 1987 to
2008. In the trade between SEA4 and EA5 over the period 1987 to 2000, for
instance, 26.7% of SEA4 total exports to EA5, along with 34.6% of EA5 total exports
to SEA4, are categorized as vertical trade. Of total Chinese exports to EA5 during
the period 1987 to 2000, vertical and processing trade combined account for 43%.
Overall, the findings are consistent with those estimates of Ando (2006), Koopman et
3 The share of upstream and midstream outputs combined in total exports is inferred to be 0.822 for EA5 and
SEA4, respectively, and 0.566 for China. The latter is in line with the estimates of Koopman et al. (2010) shown
in Table 1a, while the former presumes a value slightly higher than their estimates.
18
al. (2010), and Dean et al. (2011) in that East Asia as a region is densely intertwined
through vertical and processing trade as the corollary of vertical fragmentation of
sequential production chains across regions.
--- [INSERT TABLE 5 HERE] ---
3.4 Source of fluctuations
An interesting question appears when witnessing such tied vertical and
sequential linkage: to what extent then the foreign shocks influence home
macroeconomic fluctuations? We address this question by looking into the ability of
each structural shock in accounting for the macroeconomic volatility over the short,
medium, and long run. Table 6 reports the results of variance decomposition for
three selective macroeconomic variables, namely, GDP, CPI inflation, and short-term
nominal interest rate. Detailed findings are available upon request.
--- [INSERT TABLE 6 HERE] ---
Several worth-discussing points emerge.
Firstly, foreign influences on home macroeconomic fluctuation are substantial,
though in different ways across regions, over the short, medium, and long run. For
instance, shocks to EA5 intermediate export price mark-up and trade cost combined
can account for an approximate 37% of SEA4 GDP volatility contemporaneously, 33%
after four quarters, and 34% over the long run. Secondly, in the presence of vertical
specialization, foreign trade cost is too important to be ignored for GDP volatility,
particularly over the short and medium run. In the estimated SEA4-EA5 model,
foreign trade cost shock per se can account for 28.8% and 15.8% of the
contemporaneous GDP fluctuation in SEA4 and EA5, respectively. The fraction is
even higher in the estimated CN-EA5 model, of which 65% and 25% of CN and EA5
GDP volatility, respectively, can be attributed to disturbance on foreign trade cost.
This observation is certainly in line with the consensus that back-and-forth trade can
19
propagate even a small shock to trade cost (Yi, 2003). Thirdly, in contrast to Kim and
Lee (2008) which found a dominant role of technology and labor supply shocks in
explaining the cross-country output fluctuation, the influence of total factor
productivity (TFP) is largely confined as domestic source of fluctuation, and is
relatively unimportant when compared to investment-specific technology (IST) shock
and preference shock. It is the shock to export price – both intermediates and final
goods – markup and trade cost that matters for international macroeconomic
fluctuations.
But at CPI inflation front, fourthly, home TFP shock together with preference
shock is critically important. For instance, in the estimated SEA4-EA5 model, home
TFP shock can explain approximately 42% and 72% of contemporaneous and long-
run CPI inflation fluctuation in SEA4. The role of EA5’s own TFP shock on her CPI
inflation volatility is equally substantial. And such a role of home TFP shock can be
observed in other two estimated models, too. With respect to the foreign source of
shock, interestingly, the U.S interest rate shock has been the most common
influential foreign disturbance across the estimated models. We conjecture that
dollar pricing in trade has laid the ground for the non-trivial role of U.S interest rate in
that changes in the U.S Federal Funds rate are transmitted through nominal
exchange rates into firms’ pricing decision, and are propagated by multiple
sequential production chains.
Lastly, we find that policy interest rate in each estimated model largely
responds to domestic shocks. Shock to the U.S interest rate has been the only
common source of foreign shock across models. In the estimated CN-EA5 model
particularly, the transmission of changes in the U.S Federal Funds rate into interest
rates in China and East Asia economies is large not only contemporaneously but
also in the long run (Edwards, 2010).
20
3.5 Deciphering the influence of China
To regional economies, of which production and trade have been vertically the
most integrated in the world, the emergence of China has posed both opportunity
and threat. On the one hand, rapid expansion of China’s processing exports implies
greater demand for parts and components from its Asian neighbors. Ianchovichina
and Walmsley (2005), for instance, calibrate and simulate on a multicountry and
multisector model, and show that China’s WTO accession has crowded in Japan and
the newly industrialized economies in East Asia that supply materials to China.
Eichengreen et al. (2007) also find that China’s growth is beneficial to advanced East
Asian economies exporting capital-intensive goods but not to developing Southeast
Asian countries exporting labor-intensive goods (see also Park and Shin, 2010;
Haltmaier et al., 2007; and Roland-Holst and Weiss, 2004. Greenaway et al. 2008
argue the reverse).
Table 5 shows the influence of China especially in the aftermath of accession to
World Trade Organization in late 2001. Most dramatic change is the trade interaction
between China and the advanced East Asian economies. Prior to Chinese WTO’s
accession, both regions vertically specialize in midstream production with higher
vertical trade as a share of total exports (0.241 for China and 0.228 for EA5).
However, accession to World Trade Organization has tremendously overhauled
regional production network by repositioning China as the destination for final
assembly of parts and components shipped from the advanced East Asian neighbors
to produce and export final goods4. Processing trade accounts for larger share of
Chinese total exports to East Asia (0.377), while witnessing a rising share of vertical
trade in EA5 total exports to China (0.400).
4 See, for instance, Athukorala and Menon (2010), Athukorala and Yamashita (2009), and Hayakawa (2007).
21
Although the developing Southeast Asia remains specializing in midstream
production, the share of vertical trade in SEA4 total exports to China has been falling
drastically from 0.401 in pre China’s WTO accession to 0.251 in post China’s WTO
accession, alongside the strengthening share of processing trade in China
throughout the periods. One interesting fact to note is the developing Southeast Asia
is indeed vertically bonded more closely to the advanced East Asia rather than China
in the aftermath of China’s WTO accession.
Of question is in what way China would have affected the East and Southeast
Asia? Figure 3 depicts the dynamic responses of East and Southeast Asia, in (a) and
(b), respectively, toward one percent increase in China’s final goods export price
markup. The responses are obviously different. For China and East Asia which
complement each other in that East Asia engages in vertical trade (with nearly zero
processing trade) vis-à-vis the processing China, the dynamic responses of both
regions seems synchronized, despite the fact that the shock is asymmetric.
--- [INSERT FIGURE 3] ---
The intuition is straightforward. Thanks to the higher final goods export price
inflation, East Asian consumption is redirected from import toward home-produced
final goods. Simply, the demand for Chinese processing export will fall. The resultant
contraction in China’s downstream production in turn transmits into East Asia
through declining demand for intermediates imported from East Asia. The
repercussion effect is channeled through multiple sequential production chains.
Gross domestic product in both regions fall in consequence of the trade
complementary embodied in the vertical-processing trade structure.
However, tale is different for the case of China and Southeast Asia. There
exists not only trade complementarity between China and Southeast Asia. With
rising share of processing trade in total exports, Southeast Asia also competes head-
22
to-head with China in final goods market. And this does matter. Positive shock to
China’s final goods export price markup on the one hand makes expensive China’s
exports unfavorable, and on the other hand induces Chinese supply of final goods
toward more profitable export market, which, in turn, fuels higher price of final goods
in China’s home market. As a result, the Chinese expenditure is switched toward
imported final goods from Southeast Asia. Adverse response of China’s GDP is thus
accompanied with favorable response of Southeast Asian GDP.
4. Can Asia gain from monetary policy coordination?
We turn to an important normative question: can Asia gain from monetary
policy coordination given the growing presence of vertical and processing trade?
Speaking differently, would the lack of international monetary policy coordination
result in substantive welfare lose? The influential Obstfeld and Rogoff (2002) argue
that under plausible assumption the lack of cooperation in rule setting is of second
order. Clarida et al. (2001, 2002) argue that the policy problem for central bank of a
small open economy is isomorphic to the one it would face under closed economy.
Openness does not distinct optimal monetary policy of an open economy from that of
closed economy. Benigno and Benigno (2008) further show that the allocation under
optimal cooperative solution can be implemented through inflation-targeting regime
in such a way that each monetary authority minimizes a quadratic loss function that
targets only domestic GDP inflation and the output gap.
A worth-mentioning yet unsurprising feature of this literature is that production
is modelled either as a process with single stage or two vertically and sequentially
not connected stages 5 . In this paper, we propose a welfare criterion which
5 There are few exceptions which deserve more discussion. See, for instance, Shi and Xu (2007), and Berger
(2006). Petrella and Stantoro (2011) and Strum (2009) consider the importance of geographically defragmented
input-output interaction in optimal monetary policy.
23
characterizes the cooperative allocation for an open economy engaging in vertical
and processing trade. Following the tradition of New Keynesian literature as in
Woodford (2003), we derive a quadratic approximation of the utility-based welfare
criterion around the non-distorted global maximum utility by taking into account all
the resource constraints. We obtain the following loss function
Ð = bt Ñj�r
�st>�� − �gËÂ
�� A
of which
bt Ñj�r
�st>�� − �gËÂ
�� A
= −12 �> };�eΔA~e,�� + 'ℒ�e∗��e∗ /~�e,�∗� + 'ℒ[e∗�[e∗ /~[e,�∗� − 'ℒÔ�Ô/~Ô,�� + >ℒ|�|A ~|,�� %
−����� " D�Õ�Ö "� ×�� + �Ê
®Ù"�� Δ��®{�� + �ÚÖ"� Û��E − �Ê®Ù"�� ��l�
���" ÜÝ��� + +�Ω ßÊ"� à�� − ℱ� +
C. �. � + ã�‖å‖[� (41)
where × is real exchange rate, ÜÝ�� = Ü�� − ³� − Bæç���, and
ℱ� = ����� " Δ�� DΘ∗�Û�∗� + �Ê∗Ê ϱ∗"� ��∗� + (Ω∗ �ß∗
Ê∗" �Ê∗Ê ".� à�∗�E (42)
ℒ�e∗ = ;�∗Δ D 1 − }1 + Û + Û∗ + >�∗� A > }K[∗1 + Û∗A + >à∗�∗A >�∗� AE ℒ[e∗ = ;[∗Δ >�∗� A > 1 − }1 + Û∗A
ℒÔ = ;�K[Δ D} + >�∗� A > 1 − }1 + Û∗A + à�E ℒ| = >�®ÙA
�� D1 − B + 1 + �2 E ;|
Γ = ë�}, �, V, K�∗, K[∗, K�, K[, Û, Û∗� Θ = ë�}, K�∗ , K[∗, K�, K[, Û, Û∗� Θ∗ = ë�}, K�∗, K[∗, K�, K[, Û, Û∗�
24
Ω = K�∗K[1 + Û + Û∗ − K��−1K[� − K[
Ω∗ = K�∗�1 − K[∗� + K[∗�1 − K��1 + Û∗
ϱ∗ = }ìK�∗�1 − K[∗� + K[∗ − K�K[∗í + �1 − }�ì1 − K��1 − K[� − K[í1 + Û∗ + �1 − }�K�∗K[1 + Û + 2Û∗ Δ= }�1 − K��1 − K[� − K[� + �1 − }�ìK�∗�1 − K[∗� + K[∗ − K�K[∗í + }K�K[∗1 + Û + Û∗
What is the welfare-maximizing monetary policy rule? We consider three
games:
(i) Each central bank stabilizes only domestic variables, given foreign
monetary policy action.
(ii) Each central bank stabilizes both domestic and foreign variables, given
foreign monetary policy action.
(iii) A coordinated regime, in which there is a supranational monetary authority
maximize the GDP-weighted joint welfare function, Ðî = �Ð + �1 − ��Ð∗. [TO BE ADDED]
5. Conclusion
[TO BE ADDED]
Reference Amador, J., Cabral, S. (2009). Vertical specialization across the world: a relative
measure. The North American Journal of Economics and Finance 20(3), 267-280. Ando, M. (2006). Fragmentation and vertical intra-industry trade in East Asia. The
North American Journal of Economics and Finance 17, 257-281. Athukorala, P-C. (2007). The rise of China and East Asian export performance: is the
rowding-out fear warranted? World Economy. 32(2): 234-266. Athukorala, P-C., Menon, J. (2010). Global production sharing, trade patterns, and
determinants of trade flows in East Asia. ADB Working Paper Series on Regional Economic Integration No. 41. Asian Development Bank.
Athukorala, P., Yamashita, N. (2009). Global production sharing and Sino-US trade relations. China and World Economy. 17(3): 39-56.
Benigno, G., Benigno, P. (2008). Implementing international monetary cooperation through inflation targeting. Macroeconomic Dynamics 12, 45-59.
Berger, W. (2006). International interdependence and the welfare effects of monetary policy. International Review of Economics and Finance 15, 399-416.
25
Clarida, R., Gali, J., Gertler, M. (2001). Optimal monetary policy in open versus closed economies: an integrated approach. American Economic Review Papers and Proceedings 91(2), 248-252.
Clarida, R., Gali, J., Gertler, M. (2002). A simple framework for international monetary policy analysis. Journal of Monetary Economics 49, 879-904.
Dean, J.M., Fung, K.C., Wang, Z. (2011). Measuring vertical specialization: the case of China. Review of International Economics 19(4), 609-625.
Edwards, S. (2010). The international transmission of interest rate shocks: the Federal Reserve and emerging markets in Latin America and Asia. Journal of International Money and Finance 29(4), 685-703.
Eichengreen, B., Rhee, Y., Tong, H. (2007). China and the exports of other Asian countries. Review of World Economics. 143(2), 201-226.
Feenstra, R.C. (1998). Integration of trade and disintegration of production in the global economy. Journal of Economic Perspectives 12(4), 31-50.
Greenaway, D., Mahabir, A., Milner, C. (2008). Has China displaced other Asian countries’ exports? China Economic Review. 19, 152-169.
Haltmaier, J.T., Ahmed, S., Coulibaly, B., Knippenberg, R. (2007). The role of China in Asia: engine, conduit or streamroller? International Finance Discussion Papers No. 904, Board of Governors of the Federal Reserve System.
Obstfeld, M., Rogoff, K. (2002). Global implications of self-oriented national monetary rules. Quarterly Journal of Economics (May), 503-535.
Hayakawa, K., (2007). Growth of intermediate goods trade in East Asia. Pacific Economic Review. 12(4): 511-523.
Hummels, D., Ishii, J., Yi, K-M. (2001). The nature and growth of vertical specialization in world trade. Journal of International Economics 54, 75-96.
Ianchovichina, E., Walmsley, T. (2005). Impact of China’s WTO accession on East Asia. Contemparary Economic Policy. 23(2): 261-277.
Kim, S., Lee, J-W. (2008). International Macroeconomic Fluctuations: a new open economy macroeconomics interpretation. Working Papers 232008, Hong Kong Institute for Monetary Research.
Kim, S., Lee, J-W, Park, C-Y. (2011). Ties binding Asia, Europe and the USA. China & World Economy 19(1), 24-46.
Koopman, R., Powers, W., Wang, Z., Wei, S.-J., (2010). Give credit where credit is due: tracing value added in global production chains. NBER Working Paper No. 16426. NBER, Cambridge.
Park, D., Shin, K. (2010). Can trade with the People’s Republic of China be an engine of growth for Developing Asia? Asian Development Review 27(1), 160-181.
Roland-Holst, D., Weiss, J. (2004). ASEAN and China: export rivals or partners in regional growth?
Petrella, I., Santoro, E. (2011). Input-output interactions and optimal monetary policy. Journal of Economic Dynamics and Control 35, 1817-1830.
Sawyer, W.C., Sprinkle, R.L., Tochkov, K. (2010). Patterns and determinants of intra-industry trade in Asia. Journal of Asian Economics 21(5), 485-493.
Shi, K., Xu, J. (2007). Optimal monetary policy with vertical production and trade. Review of International Economics 15(3), 514-537.
Strum, B.E., (2009). Monetary policy in a forward-looking input-output economy. Journal of Money, Credit and Banking 41(4), 619-650.
Wong, C.Y., Eng, Y.K. (2011). International business cycle comovement and vertical specialization reconsidered in multistage Bayesian DSGE model. FREIT Working Paper No. 396.
26
Table 1. The priors for parameters and shocks
Prior distribution Probability distribution
function
Mean Standard deviation
Parameters
Risk aversion coefficient � Uniform 1 0.577
Reciprocal of wage elasticity of labor supply �< Gamma 2 1.000
Habit persistence � Beta 0.7 0.100
Forward looking-ness of investment decision Λ Uniform 0.5 0.289
Els btw. home and imported intermediate goods V Normal 1.5 0.500
Home bias in consumption } Beta 0.7 0.100
Share of imported materials at intermediate production K�
Uniform 0.5 0.289
Share of imported intermediate goods at final production K[
Uniform 0.5 0.289
Employment indexation }| Uniform 0.5 0.289
Producer price indexation }�� Uniform 0.5 0.289
Final goods price indexation }�[ Uniform 0.5 0.289
Intermediate export price indexation }��∗ Uniform 0.5 0.289
Final goods export price indexation }�[∗ Uniform 0.5 0.289
Employment stickiness ]h Uniform 0.75 0.144
Producer price stickiness ]�� Uniform 0.75 0.144
Final goods price stickiness ]�[ Uniform 0.75 0.144
Intermediate export price stickiness ]��∗ Uniform 0.75 0.144
Final export price stickiness ]�[∗ Uniform 0.75 0.144
Policy inertia ¯� Beta 0.7 0.100
Policy response to inflation °± Gamma 1.5 1.000
Policy response to GDP fluctuation °® Gamma 0.125 0.050
Policy response to exchange rate variability °∆w Gamma 0.5 0.100
TFP shock persistence ¯Ë Beta 0.8 0.100
IST shock persistence ¯� Beta 0.7 0.100
Shocks
Total factor productivity �Ë Uniform 0.5 0.289
Investment-specific technology �� Uniform 0.5 0.289
Labor supply �^ Uniform 0.5 0.289
Preference �Ê Uniform 0.5 0.289
Producer price markup �±QÁ Uniform 0.5 0.289
Final goods price markup �±\Á Uniform 0.5 0.289
Intermediate export price markup �±QÁ∗ Uniform 0.5 0.289
Final export price markup �±\Á∗ Uniform 0.5 0.289
Transportation cost �ï Uniform 0.5 0.289
Monetary policy �? Uniform 0.5 0.289
UIPC �ΠUniform 0.5 0.289
U.S monetary policy �ÔÔ? Uniform 0.5 0.289
27
Table 2. Selected posterior distributions for SEA4-EA5 model
1987Q1-2000Q4 2001Q1-2008Q4
Southeast Asia East Asia Southeast Asia East Asia Mode 5% Mean 95% Mode 5% Mean 95% Mode 5% Mean 95% Mode 5% Mean 95%
Parameters
� 0.392 0.291 0.352 0.416 0.392 0.291 0.352 0.416 0.503 0.370 0.481 0.585 0.503 0.370 0.481 0.585 } 0.758 0.725 0.773 0.821 0.941 0.925 0.941 0.957 0.659 0.614 0.667 0.746 0.889 0.861 0.889 0.920 V 1.548 1.483 1.551 1.611 1.548 1.483 1.551 1.611 1.577 1.521 1.592 1.663 1.577 1.521 1.592 1.663 Λ 1.000 0.976 0.946 1.000 1.000 0.976 0.946 1.000 0.963 0.926 0.965 1.000 0.963 0.926 0.965 1.000 K� 0.620 0.527 0.649 0.745 0.854 0.729 0.843 0.995 1.000 0.814 0.921 1.000 0.577 0.511 0.622 0.743 K[ 0.173 0.087 0.148 0.217 0.786 0.687 0.811 0.943 0.510 0.328 0.511 0.667 0.768 0.621 0.735 0.858 ]�� 0.620 0.592 0.619 0.646 0.871 0.843 0.867 0.890 0.602 0.572 0.621 0.685 0.691 0.634 0.680 0.722 ]�[ 0.776 0.750 0.777 0.805 0.943 0.936 0.944 0.952 0.851 0.808 0.841 0.873 0.910 0.892 0.914 0.933 ]��∗ 0.955 0.932 0.946 0.960 0.587 0.520 0.590 0.662 0.703 0.640 0.737 0.860 0.734 0.650 0.715 0.780 ]�[∗ 0.639 0.604 0.637 0.678 0.699 0.671 0.699 0.727 0.652 0.605 0.644 0.683 0.770 0.740 0.772 0.802 ¯Ë 0.809 0.760 0.820 0.882 0.947 0.936 0.944 0.952 0.890 0.855 0.897 0.937 0.902 0.840 0.886 0.944 ¯� 0.799 0.739 0.771 0.801 0.640 0.559 0.630 0.698 0.705 0.628 0.662 0.698 0.597 0.556 0.597 0.645
Shocks �Ë 0.047 0.034 0.055 0.074 0.030 0.025 0.036 0.047 0.038 0.030 0.043 0.056 0.024 0.019 0.030 0.041 �� 0.023 0.021 0.028 0.034 0.027 0.019 0.029 0.038 0.017 0.016 0.022 0.027 0.030 0.023 0.031 0.040 �q 0.059 0.049 0.061 0.073 0.033 0.027 0.033 0.039 0.036 0.029 0.038 0.047 0.017 0.029 0.038 0.047 �±QÁ 0.157 0.125 0.167 0.214 0.659 0.447 0.643 0.873 0.123 0.081 0.142 0.220 0.104 0.060 0.099 0.136 �±\Á 0.192 0.142 0.219 0.318 0.866 0.762 0.887 1.000 0.296 0.000 0.192 0.393 0.427 0.342 0.479 0.618 �±QÁ∗ 1.000 0.805 0.894 1.000 0.633 0.610 0.687 0.773 0.279 0.223 0.348 0.480 0.664 0.429 0.595 0.752 �±\Á∗ 0.440 0.374 0.472 0.571 0.657 0.573 0.709 0.839 0.261 0.198 0.271 0.339 0.364 0.321 0.429 0.537 �? 0.014 0.011 0.014 0.018 0.004 0.004 0.005 0.006 0.011 0.008 0.011 0.015 0.003 0.003 0.003 0.004
Notes: The posterior distribution is obtained using the Metropolis-Hastings sampling algorithm based on 4 parallel chains of 50,000 draws, of which the first half was discarded as burn-in. The average acceptance rate is 0.238 for estimation of subsample 1987Q1-2000Q4, and 0.272 for 2001Q1-2008Q4. We impose identical posteriors for �, V, and Λ across regions.
28
Table 3. Selected posterior distributions for CN-EA5 model
1987Q1-2000Q4 2001Q1-2008Q4
China East Asia China East Asia Mode 5% Mean 95% Mode 5% Mean 95% Mode 5% Mean 95% Mode 5% Mean 95%
Parameters
� 0.350 0.297 0.352 0.414 0.350 0.297 0.352 0.414 0.349 0.285 0.344 0.404 0.349 0.285 0.344 0.404 } 0.763 0.720 0.751 0.778 0.888 0.851 0.880 0.904 0.457 0.410 0.444 0.478 0.706 0.671 0.702 0.728 V 1.409 1.384 1.428 1.470 1.409 1.384 1.428 1.470 1.557 1.484 1.528 1.576 1.557 1.484 1.528 1.576 Λ 1.000 0.968 0.986 1.000 1.000 0.968 0.986 1.000 0.997 0.953 0.977 1.000 0.997 0.953 0.977 1.000 K� 0.841 0.749 0.851 0.968 0.523 0.455 0.555 0.712 0.490 0.405 0.519 0.623 1.000 0.941 0.973 1.000 K[ 0.439 0.370 0.436 0.503 0.762 0.715 0.805 0.901 0.841 0.775 0.869 0.969 0.391 0.346 0.406 0.461 ]�� 0.591 0.578 0.610 0.642 0.845 0.829 0.848 0.865 0.670 0.632 0.663 0.691 0.663 0.630 0.670 0.710 ]�[ 0.899 0.892 0.903 0.913 0.932 0.926 0.935 0.944 0.940 0.936 0.950 0.964 0.917 0.907 0.922 0.933 ]��∗ 0.938 0.930 0.942 0.954 0.769 0.691 0.738 0.780 0.740 0.711 0.780 0.850 0.925 0.920 0.940 0.962 ]�[∗ 0.726 0.707 0.732 0.753 0.555 0.532 0.558 0.588 0.648 0.619 0.653 0.700 0.729 0.719 0.739 0.758 ¯Ë 0.940 0.934 0.939 0.945 0.940 0.934 0.942 0.950 0.940 0.930 0.938 0.946 0.713 0.681 0.725 0.771 ¯� 0.644 0.667 0.702 0.734 0.681 0.642 0.676 0.713 0.752 0.730 0.768 0.812 0.659 0.644 0.664 0.682
Shocks �Ë 0.102 0.078 0.102 0.124 0.032 0.024 0.035 0.044 0.039 0.030 0.041 0.054 0.048 0.030 0.049 0.067 �� 0.037 0.022 0.029 0.035 0.022 0.019 0.024 0.029 0.004 0.002 0.003 0.004 0.026 0.020 0.026 0.032 �q 0.077 0.067 0.080 0.092 0.037 0.033 0.039 0.046 0.025 0.019 0.025 0.031 0.021 0.014 0.023 0.031 �±QÁ 0.129 0.116 0.154 0.197 0.419 0.303 0.411 0.506 0.050 0.029 0.052 0.073 0.106 0.068 0.123 0.175 �±\Á 0.864 0.857 0.932 1.000 0.559 0.496 0.632 0.739 0.353 0.358 0.517 0.677 0.609 0.527 0.669 0.780 �±QÁ∗ 0.635 0.637 0.712 0.785 1.000 0.904 0.957 1.000 0.580 0.448 0.570 0.706 0.209 0.160 0.241 0.312 �±\Á∗ 0.670 0.540 0.631 0.721 0.202 0.157 0.202 0.243 0.181 0.131 0.209 0.288 0.118 0.100 0.166 0.228 �? 0.013 0.010 0.013 0.015 0.005 0.004 0.005 0.006 0.005 0.004 0.006 0.007 0.004 0.003 0.005 0.006
Notes: The posterior distribution is obtained using the Metropolis-Hastings sampling algorithm based on 4 parallel chains of 50,000 draws, of which the first half was discarded as burn-in. The average acceptance rate is 0.237 for estimation of subsample 1987Q1-2000Q4, and 0.346 for 2001Q1-2008Q4. We impose identical posteriors for �, V, and Λ across regions.
29
Table 4. Selected posterior distributions for CN-SEA4 model
1987Q1-2000Q4 2001Q1-2008Q4
China Southeast Asia China Southeast Asia Mode 5% Mean 95% Mode 5% Mean 95% Mode 5% Mean 95% Mode 5% Mean 95%
Parameters
� 0.578 0.465 0.544 0.621 0.578 0.465 0.544 0.621 0.916 0.812 1.020 1.209 0.916 0.812 1.020 1.209 } 0.734 0.657 0.697 0.737 0.677 0.626 0.664 0.712 0.844 0.808 0.862 0.902 0.603 0.579 0.618 0.663 V 1.417 1.355 1.424 1.484 1.417 1.355 1.424 1.484 1.592 1.519 1.559 1.603 1.592 1.519 1.559 1.603 Λ 0.860 0.821 0.884 0.949 0.860 0.821 0.884 0.949 0.927 0.861 0.920 0.991 0.927 0.861 0.920 0.991 K� 0.828 0.816 0.898 1.000 1.000 0.947 0.975 1.000 0.512 0.377 0.518 0.709 0.732 0.424 0.610 0.771 K[ 0.896 0.818 0.899 1.000 0.685 0.625 0.684 0.744 1.000 0.901 0.949 1.000 0.637 0.597 0.752 0.919 ]�� 0.605 0.568 0.589 0.611 0.739 0.707 0.732 0.754 0.656 0.623 0.657 0.694 0.567 0.513 0.563 0.611 ]�[ 0.909 0.890 0.902 0.915 0.864 0.848 0.864 0.881 0.935 0.927 0.941 0.954 0.856 0.835 0.855 0.874 ]��∗ 0.947 0.932 0.950 0.967 0.909 0.889 0.913 0.938 0.745 0.664 0.744 0.813 0.648 0.523 0.643 0.732 ]�[∗ 0.663 0.643 0.671 0.701 0.594 0.571 0.598 0.625 0.715 0.677 0.707 0.739 0.555 0.500 0.522 0.546 ¯Ë 0.945 0.933 0.944 0.955 0.949 0.930 0.944 0.959 0.852 0.838 0.887 0.936 0.894 0.839 0.884 0.931 ¯� 0.707 0.680 0.725 0.765 0.687 0.664 0.710 0.755 0.599 0.556 0.612 0.680 0.678 0.561 0.616 0.676
Shocks �Ë 0.055 0.044 0.058 0.070 0.054 0.039 0.057 0.075 0.025 0.014 0.022 0.030 0.041 0.030 0.046 0.061 �� 0.028 0.019 0.027 0.034 0.042 0.027 0.038 0.049 0.008 0.005 0.008 0.012 0.019 0.017 0.026 0.034 �q 0.071 0.061 0.074 0.088 0.066 0.057 0.071 0.083 0.029 0.022 0.030 0.037 0.060 0.022 0.030 0.037 �±QÁ 0.126 0.091 0.119 0.147 0.635 0.476 0.576 0.682 0.062 0.042 0.065 0.086 0.089 0.056 0.098 0.138 �±\Á 0.736 0.637 0.742 0.853 0.932 0.728 0.853 0.993 0.972 0.727 0.855 0.994 0.996 0.901 0.956 1.000 �±QÁ∗ 0.790 0.733 0.871 1.000 0.727 0.654 0.769 0.897 0.389 0.114 0.224 0.331 0.490 0.415 0.573 0.742 �±\Á∗ 0.399 0.332 0.413 0.487 0.248 0.191 0.250 0.305 0.311 0.253 0.326 0.407 0.059 0.029 0.056 0.082 �? 0.008 0.006 0.008 0.010 0.033 0.024 0.032 0.039 0.004 0.002 0.004 0.006 0.012 0.009 0.013 0.017
Notes: The posterior distribution is obtained using the Metropolis-Hastings sampling algorithm based on 4 parallel chains of 50,000 draws, of which the first half was discarded as burn-in. The average acceptance rate is 0.346 for estimation of subsample 1987Q1-2000Q4, and 0.251 for 2001Q1-2008Q4. We impose identical posteriors for �, V, and Λ across regions.
30
Table 5. Measuring vertical specialization of total export
Pre China's WTO accession Post China's WTO accession Share of imported
intermediate inputs Export share
Vertical specialization
Share of imported intermediate
inputs Export share Vertical
specialization
Model Mid-
stream Down-stream
Mid-stream
Down-stream
Vertical trade
Processing trade Index
Mid-stream
Down-stream
Mid-stream
Down-stream
Vertical trade
Processing trade Index
SEA4-EA5
SEA 0.649 0.148 0.411 0.178 0.267 0.026 0.293 0.921 0.511 0.411 0.178 0.379 0.091 0.469
EA 0.843 0.811 0.411 0.178 0.346 0.144 0.491 0.622 0.735 0.411 0.178 0.256 0.131 0.386
CN-EA5
China 0.851 0.436 0.283 0.434 0.241 0.189 0.430 0.519 0.869 0.283 0.434 0.147 0.377 0.524
EA 0.555 0.805 0.411 0.178 0.228 0.143 0.371 0.973 0.406 0.411 0.178 0.400 0.072 0.472
CN-SEA4
China 0.898 0.899 0.283 0.434 0.254 0.390 0.644 0.518 0.949 0.283 0.434 0.147 0.412 0.558
SEA 0.975 0.684 0.411 0.178 0.401 0.122 0.522 0.610 0.752 0.411 0.178 0.251 0.134 0.385 Notes: The formula for computing the vertical specialization of total export for country � is given by °u- = K� � £Â�f?�Q∑ £Â�f?�Á\Q "ÅÆÆÆÇÆÆÆÈ
Éh?�-ÊËÌ�?ËÍh+ K[ � £Â�f?�\∑ £Â�f?�Á\Q "ÅÆÆÆÇÆÆÆÈ
�?fÊhÎÎ-^Ï�?ËÍh. The share of midstream and downstream output in total export is inferred from Kim et al. (2011).
SEA4 consists of Indonesia, Malaysia, Thailand, and the Philippines, and EA5 consists of Japan, the Republic of Korea, Hong Kong, Taiwan, and Singapore. All are weighted by time-varying total trade share.
31
Table 6. Variance decomposition, 2001Q1-2008Q4: foreign influences are important in different ways
(a) GDP
SEA4 EA5 Home Foreign Home Foreign
t 7Ê 7±QÁ∗ 7±\Á∗ 7±Qð 7ï∗ ñò∗ ñó∗ ñô∗ ñõ 0 19.37 6.33 12.5 8.57 28.77 3.02 36.2 31.8 15.8 2 18.17 8.9 15.0 13.7 18.93 5.4 41.2 30.3 10.7 4 14.31 11.67 15.66 20.7 12.74 9.5 41.3 26.8 8.27 8 11.24 13.69 14.34 25.45 9.86 12.9 37.6 24.0 7.34 16 10.47 14.38 13.56 25.37 9.2 12.6 36.1 22.6 6.82 ∞ 10.37 14.33 13.44 25.13 9.11 12.4 35.9 22.3 6.73
CN EA5
Home Foreign Home Foreign
t 7Ë 7±QÁ∗ 7±\Á∗ 7ï∗ 7±\ð∗ ñó∗ ñô∗ ñ÷øù ñò ñ÷úû∗ ñõ 0 0.02 1.4 5.85 65.07 6.51 28.2 33.5 0.21 0.01 0.12 25.3
2 0.05 3.89 9.12 51.38 10.24 32.3 33.8 0.46 0.02 0.27 18.1
4 0.12 9.4 13.55 37.12 13.15 32.6 30.9 1.25 0.06 1.9 14.6
8 0.24 13.76 19.55 25.78 11.95 28.6 26.8 3.31 0.2 9.18 12.6
16 1.32 12.38 21.04 21.16 9.78 25.7 24.1 6.4 1.53 11.92 11.3 ∞ 18.12 4.81 6.3 7.86 2.85 16.6 15.5 23.5 14.3 7.85 7.21
CN SEA4
Home Foreign Home Foreign
t 7±QÁ∗ 7Ë∗ 7±Qð 7±\ð∗ ñò∗ ñô∗ ñ÷øù ñõ 0 2.29 0.47 28.44 34.33 1.91 13.1 63 14.6
2 4.61 0.38 36.43 31.13 2.86 11.4 69 8.86
4 9.8 0.29 43.77 24.32 5.28 9.27 71.1 6.05
8 17.3 2.36 43.02 16.73 10 7.87 68.1 4.93
16 20.44 8.26 36.22 12.06 14.54 7.22 63.5 4.5 ∞ 20.22 11.95 33.42 10.77 15.73 7.07 62.2 4.41
32
(b) CPI Inflation
SEA4 EA5
Home Foreign Home Foreign
t 7Ë 7Ê 7±\ð 7ÔÔ? ñò∗ ñó∗ ñ÷úû∗ ñùùü 0 41.63 21.25 9.37 10.3 32.01 10.54 16.21 22.09
2 51.29 16.28 8.04 8.7 40.55 8.84 15.59 14.72
4 61.17 12.26 6.01 7.46 51.08 7.24 11.11 9.7
8 68.33 9.56 4.71 6.55 60.1 5.93 7.86 7.07
16 71.54 8.39 4.14 6.02 64.95 5.37 6.87 6.09 ∞ 72.14 8.19 4.04 5.89 65.82 5.25 6.67 9.91
CN EA5
Home Foreign Home Foreign
t 7Ë 7Ê 7? 7ÔÔ? ñò∗ ñ÷úû∗ ñùùü 0 34.41 12.62 9.96 26.32 10.62 26.13 28.21
2 38.28 11.44 9.03 23.88 10.61 33.26 21.25
4 45.39 9.77 7.71 20.39 15.19 31.31 18.01
8 54.22 7.8 6.16 16.37 22.47 26.55 16.5
16 61.12 6.39 5.05 13.6 24.02 26.4 15.86 ∞ 63.83 5.79 4.58 12.33 23.93 26.39 15.83
CN SEA4
Home Foreign Home
t 7Ë 7Î 7±Qð 7ÔÔ? ñò∗ ñó∗ ñô∗ 0 30.55 27.07 0.05 17.7 46.21 8.91 14.33
2 39.16 24.85 5.18 10.01 56.29 7.77 10.22
4 54.46 11.74 11.42 5.59 67.79 6.42 7.32
8 62.49 5.55 13.91 3.37 75.12 5.11 5.54
16 64.64 3.93 12.76 2.53 78.04 4.51 4.87 ∞ 64.31 3.7 12.22 2.39 78.48 4.42 4.47
33
(c) Nominal interest rate
SEA EA
Home Foreign Home Foreign
t 7Ë 7Ê 7±QÁ∗ 7±\Á∗ 7ÔÔ? ñó∗ ñô∗ ñü∗ ñùùü 0 10.23 10.77 9.74 8.01 46.41 0.13 5.55 21.13 59.02
2 8.64 13.15 9.72 6.22 47.13 2.27 7.13 22.48 56.17
4 6.56 15.36 8.09 4 44.62 9.76 8.79 21.48 46.43
8 6.52 14.98 7.63 7.18 38.91 17.29 8.98 16.78 32.13
16 7.93 14.03 9.85 14.38 36.17 17.93 8.07 13.25 24.71 ∞ 8.48 13.87 10.01 15.67 35.75 17.47 7.83 12.8 23.9
CN EA
Home Foreign Home Foreign
t 7Ë 7±QÁ∗ 7±Qð 7ÔÔ? ñò∗ ñô∗ ñü∗ ñ÷úû∗ ñùùü 0 16.6 14.12 13.43 42.9 7.23 24.41 31.23 9.86 15.13
2 15.74 15.77 16.01 39.98 14.07 20.68 24.71 17.9 10.87
4 14.53 16.74 20.09 35.36 19.61 15.87 17.39 24.24 8.19
8 13.51 15.42 23.63 31.55 18.5 12.66 13.28 22.51 11.19
16 12.19 17.43 22 28.39 17.2 10.74 11.35 22.1 15.15 ∞ 12.89 17.77 22.06 25.72 17.04 10.55 11.15 22 15.01
CN SEA
Home Foreign Home Foreign
t 7Ë 7Ê 7? 7Î 7ÔÔ? ñô∗ ñý∗ ñ÷øù ñùùü 0 2.20 17.97 14.98 38.41 19.32 8.37 37.15 22.79 17.14
2 9.25 20.74 16.63 25.11 19.66 10.3 26.36 29.11 19.48
4 20.82 19.56 15.04 15.89 17.34 11.68 20.39 30.95 20.66
8 25.83 16.67 12.53 11.83 16.91 12.15 17.98 28.61 20.57
16 23.51 14.78 11.08 10.44 17.2 10.91 15.53 33.62 17.9 ∞ 23.22 14.37 10.78 10.15 16.82 10.41 14.67 35.06 16.92
34
0 5 10 15-1
0
1GDP
0 5 10 15-1
0
1Consumption
0 5 10 15-1
0
1Export
0 5 10 15-1
0
1Import
0 5 10 15-1
0
1PPI inflation
0 5 10 15-1
0
1Export deflator inflation
0 5 10 15-1
0
1CPI inflation
0 5 10 15-1
0
1Interest rate
Data SEA-EA Model: SEA
0 5 10 15-1
0
1GDP
0 5 10 15-1
0
1Consumption
0 5 10 15-1
0
1Investment
0 5 10 15-1
0
1Exchange rate
0 5 10 15-1
0
1Export
0 5 10 15-1
0
1Import
0 5 10 15-1
0
1Export deflator inflation
0 5 10 15-1
0
1CPI inflation
data CN-EA Model:EA5
(a)
(b)
35
(c)
Fig. 1. Model fit in replicating the actual autocorrelation after China’s WTO accession
0 5 10 15-1
0
1GDP
0 5 10 15-1
0
1Consumption
0 5 10 15-1
0
1Investment
0 5 10 15-1
0
1Exchange rate
0 5 10 15-1
0
1Export
0 5 10 15-1
0
1Import
0 5 10 15-1
0
1Export deflator inflation
0 5 10 15-1
0
1CPI inflation
Data CN-SEA Model:SEA4
36
(a) CN-EA5 model
0 5 10 15 20-0.05
0
0.05GDP
0 5 10 15 20-0.01
0
0.01Hours worked
0 5 10 15 20-0.04
-0.02
0
0.02Consumption
0 5 10 15 20-4
-2
0
2x 10
-4 Investment
0 5 10 15 20-0.4
-0.2
0Export
0 5 10 15 20-0.2
0
0.2Import
0 5 10 15 20-0.1
0
0.1Nominal exchange rate
0 5 10 15 20-0.02
0
0.02
0.04CPI inflation
CN EA5
37
(b) CN-SEA4 model
Fig. 3. Dynamic response to 1% China’s final export price markup shock, 2001Q1-2008Q4
0 5 10 15 20-0.1
0
0.1GDP
0 5 10 15 20-4
-2
0
2x 10
-3 Hours worked
0 5 10 15 20-5
0
5
10x 10
-3 Consumption
0 5 10 15 20-1.5
-1
-0.5
0x 10
-3 Investment
0 5 10 15 20-0.2
0
0.2Export
0 5 10 15 20-0.02
0
0.02Import
0 5 10 15 20-0.06
-0.04
-0.02
0Nominal exchange rate
0 5 10 15 20-0.01
0
0.01
0.02CPI inflation
CN SEA4